Teaching and Teacher Education 118 (2022) 103826

Contents lists available at ScienceDirect

Teaching and Teacher Education

journal homepage: www.elsevier.com/locate/tate

Research paper

Understanding gender bias in teachers’ grading: The role of gender

stereotypical beliefs
L. Doornkamp a, *, L.D. Van der Pol a, S. Groeneveld b, J. Mesman a, J.J. Endendijk c,
M.G. Groeneveld a
a
Leiden University College, University Leiden, Turfmarkt 99, 2511DC, Den Haag, the Netherlands
b
Institute of Public Administration, University Leiden, Turfmarkt 99, 2511DC, Den Haag, the Netherlands
c
Child and Adolescent Studies, University Utrecht, Heidelberglaan 1, 3584CS, Utrecht, the Netherlands

a r t i c l e i n f o a b s t r a c t

Article history: This study experimentally tested the influence of secondary school students' gender on Dutch language
Received 5 April 2022 and math teachers' grading (n ¼ 358) and examined the role of teachers' gender and gender stereotypes
Received in revised form in gender grading bias. Results showed that grading, on average, was not gender biased. However, dif-
27 June 2022
ferences between teachers' gender grading bias were related to their gender stereotypes. Although we
Accepted 6 July 2022
found no direct effect of teachers' gender on grading bias, for math we found an indirect effect through
Available online 6 August 2022
their gender stereotypes. This study provides evidence for the role of teachers’ gender stereotypes in
gender grading bias that thus far had only been assumed.
Keywords:
Gender grading bias
© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license
Implicit gender stereotypes (http://creativecommons.org/licenses/by/4.0/).
Implicit gender-typed associations
Implicit gender-typed expectations

1. Introduction
Gender bias in teachers' grading refers to the differences in
grades between male and female students with the same level of
academic skills (Protivínský & Münich, 2018). As grades are inter-
preted as objective measures of students' level of academic skills,
gender bias in teachers' grading can have far-reaching conse-
quences. For instance, gender bias in teachers' grading was found to
be associated with students' enrollment in advanced level math
courses which has long-term implications for students’ future ca-
reers (Lavy & Sand, 2018).
Despite the multitude of studies on the topic of gender grading
bias, still little is known about the causal effect of students' gender
on teachers' grading as experimental designs to study the effect are
rare (Hanna & Linden, 2012; Hinnerich, Ho € glin, & Johannesson,
2011). Further, although scholars urged that the role of teachers'
characteristics in gender grading bias should be explored (Falch &
Naper, 2013; Lavy, 2008; Mate j
u & Smith, 2015; Protivínský &
Münich, 2018), thus far, only a few studies empirically tested the
association between teachers' gender and gender grading bias
(Lavy, 2008; Lindahl, 2016). Moreover, the role of teachers’ implicit
gender stereotypes has often been assumed, but not empirically
tested, to underly gender grading bias (Breda & Ly, 2015; Burgess &
Greaves, 2013; Gibbons & Chevalier, 2008; Lavy, 2008; Lavy & Sand,
2018; Mate j
u & Smith, 2015).
In the current study we use a field experiment among Dutch
teachers in training in the subjects Dutch language and math to
contribute to the understanding of gender grading bias. We aim to
answer the research question: to what extent is teachers' grading
gender biased and what is the role of teachers' gender and teachers'
implicit gender stereotypes in gender grading bias? By manipu-
lating gendered names on tests, we examine the extent to which
teachers' grading is influenced by students' perceived gender. We
further investigate whether gender grading bias is associated with
teachers' gender and implicit gender stereotypes and we propose
and empirically test a model that explains the potential influence of
the gender of the teacher on gender grading bias through teachers’
implicit gender-typed associations and gender-typed expectations.
In doing so, we are able to make inferences about causality, about
the role of teacher characteristics and about a mechanism under-
lying gender grading bias.

* Corresponding author.
E-mail address: l.doornkamp@luc.leidenuniv.nl (L. Doornkamp).

https://doi.org/10.1016/j.tate.2022.103826
0742-051X/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
L. Doornkamp, L.D. Van der Pol, S. Groeneveld et al. Teaching and Teacher Education 118 (2022) 103826

1.1. Gender grading bias H1. Perceived male gender of students is negatively associated
with teachers' grading.
The literature on gender grading bias consists of several studies
without a clear-cut theory that binds them together. Studies on the
1.2. The role of teachers’ gender and gender stereotypes in gender
phenomenon of gender grading bias date back to at least 1967
grading bias
(Caldwell & Hartnett, 1967). Since then, the topic has been studied
in the fields of economics, education, and sociology, each using
The role of teacher characteristics in gender grading bias has
different terminologies to refer to the same phenomenon. Gender
often been suggested in the literature (e.g., Mate j
u & Smith, 2015),
grading bias (Berg, Palmgren, & Tyrefors, 2020), gender grading gap
but only a handful of studies have empirically tested the associa-
(Falch & Naper, 2013), gender disparities in grading (Cornwell,
tions between teacher characteristics and gender grading bias
Mustard, & Van Parys, 2013), gender bias in examination (Stobart,
(Lavy, 2008; Lavy & Sand, 2018; Lindahl, 2016). Teachers’ gender,
Elwood, & Quinlan, 1992), evaluation bias (Breda & Ly, 2015),
age, experience, number of sons and daughters, and marital status
assessment bias (Lindahl, 2016), and teacher discrimination (Lavy,
were found to be related to gender grading bias, indicating that
2008) all refer to teachers' systematic over- or underassessment
gender grading bias is indeed sensitive to teacher characteristics
of a group of students based on students' gender. The literature on
(Lavy, 2008; Lavy & Sand, 2018).
gender grading bias treats gender as a binary concept as it describes
Teachers' own gender (whether they classify themselves as male
teachers' differential grading of male versus female students (e.g.,
or female) can play an important role in gender grading bias. In
Protivínský & Münich, 2018). We are aware that gender can entail
schools, some subjects are stereotyped as typically masculine (e.g.,
more than these binary categories. However, societies generally are
math) or typically feminine (e.g., languages). Although the magni-
still organized in a binary way, categorizing humans as either male
tude of the ‘math ¼ masculine’ gender stereotype can differ across
or female. This categorization lies at the basis of gender stereotypes
cultures, the content of the stereotype generally is the same (Breda,
and the accompanying gender discriminative behaviors in
Jouini, Napp, & Thebault, 2020; Nosek et al., 2009). A female
everyday life (Hyde, Bigler, Joel, Tate, & Van Anders, 2019). As we
teacher in a stereotypically masculine subject (e.g., a female math
study the role of gender stereotypes in teachers’ actual grading
teacher) may have accumulated counter-gender stereotypic expe-
behavior, we follow the gender binary approach.
riences that affect her gender bias in grading (Lavy & Sand, 2018).
The vast majority of studies on the topic used secondary data
Because of her personal experience that a woman can perform well
and relied on the comparison of students' exam scores obtained
in a subject that is stereotyped as masculine, a female math teacher
from non-blind grading (generally the teacher is the grader and is
can have weaker stereotypic ideas and beliefs regarding male and
aware of student gender) to students' exam scores obtained from
female students in math (Bussey & Bandura, 1999). These experi-
blind grading (the assessor is not aware of student gender, these
ences might even result in counter-stereotypic beliefs that might
scores are interpreted as objective measures of students’ academic
turn into counter-stereotypic behaviors (Crisp, Bache, & Maitner,
level of skills). If the gender differences in test scores awarded by
2009; Endendijk et al., 2013). In general, though, experiences
teachers (non-blind grading) were larger than the gender differ-
affect the strength of gender stereotypes but not the direction of
ences in blindly graded standardized entrance exams (Protivínský
the stereotype (i.e., gender stereotypes are flexible but remain
& Münich, 2018), central exit exams (Falch & Naper, 2013), exter-
gender stereotypic) (Charlesworth & Banaji, 2021; Xu & Meier,
nally graded state exams (Lavy, 2008), or national tests (Lindahl,
2021). Therefore, counter-stereotypic experiences are more likely
2016), these studies concluded that there is gender grading bias.
to result in weaker gender stereotypical beliefs and behaviors (for a
Generally, blind grading produces lower test scores for all students
meta-analysis see Lenton, Bruder, & Sedikides, 2009). Following
(male and female) than non-blind (teacher) grading (Hinnerich
this logic, female teachers would have weaker gender grading bias
et al., 2011). More importantly, whereas blind grading shows no
in math than male teachers, and male teachers would have weaker
or little differences in average male and female test scores, teachers
gender grading bias in languages than female teachers (Lavy, 2008).
do seem to differentiate between male and female students in their
grading, i.e., gender grading bias (e.g., Protivínský & Münich, 2018). H2a. Female math teachers display less gender grading bias than
The differences in male and female scores as a result of teachers' male math teachers.
grading is generally to the disadvantage of male students, meaning
H2b. Male Dutch language teachers display less gender grading
that male students systematically receive lower scores than female
bias than female Dutch language teachers.
students in teacher grading (Protivínský & Münich, 2018). Tens of
previous studies found evidence for this grading bias against male Indeed, teachers' gender stereotypes are presumed to underly
students across several subjects, in different countries, and for gender grading bias (Breda & Ly, 2015; Hanna & Linden, 2012;
different educational levels (for an overview see Protivínský & Hinnerich et al., 2011; Lavy, 2008; Lavy & Sand, 2018; Lindahl, 2016;
Münich, 2018). Whereas only a handful of studies found biased Mate j
u & Smith, 2015). Gender stereotypes are socially constructed
grading against female students in different subjects like math, but ideas that provide descriptions of what men and women are like,
also in humanities and English (Breda & Ly, 2015; Gibbons & and prescriptions about what men and women should be like
Chevalier, 2008; Lavy & Sand, 2018), or did not find gender (Ellemers, 2018). The content of gender stereotypes is assumed to
grading bias at all (Hanna & Linden, 2012; Hinnerich et al., 2011). be similar across different cultures, however it should be noted that
Notably, studies that did not find gender bias in grading were the recent research demonstrated that culture can shape how men and
only few studies that have experimentally tested the influence of women are perceived based on gender stereotypes (Obioma,
students' gender on teachers' grading (Hanna & Linden, 2012; Hentschel, & Hernandez Bark, 2021). Obioma et al. (2021)
Hinnerich et al., 2011). Hence, little is known about whether there showed that the content of the gender stereotype ‘men are
is a causal relation between students' gender and teachers' grading. agentic, women are communal’ is similar in Germany and Nigeria.
Insight in cause-effect provides the basis for methods to reduce However, in their study, in Germany men were perceived as more
grading bias. In all, most of the evidence on the topic of gender agentic than women while men and women were perceived as
grading bias points in the direction of a grading bias in teachers’ equally communal. In Nigeria, men and women were perceived as
grading against male students, we therefore hypothesize: equally agentic while women were perceived as more communal
than men (Obioma et al., 2021). Notwithstanding differences in
2
L. Doornkamp, L.D. Van der Pol, S. Groeneveld et al.
perception based on gender stereotypes, individuals who have
strong gender stereotypes make a strong distinction between men
and women and between typically masculine and typically femi-
nine attributes and behaviors (i.e., gender-typed individuals/
gender-schematic individuals) (Bem, 1981). When confronted with
someone else's behavior, these individuals are likely to perceive the
behavior in terms of congruence with their ideas about and ex-
pectations of males and females (Bem, 1981). Behavior in line with
these ideas and expectations are rewarded, whereas incongruent
behavior is penalized. Consequently, the distinction between males
and females based on gender stereotypes can result in different
evaluations of the same behavior of males and females (e.g., gender
grading bias) (e.g., Burgess & Borgida, 1999). Following this logic,
teachers who have strong gender stereotypes are likely to make a
strong distinction about what is typical and appropriate for male
and female students. As a result, the same behavior in males and
females is evaluated differently which makes gender bias in
grading more likely.
An individual can have implicit and explicit gender stereotypes.
Implicit gender stereotypes refer to the automatic and less
controllable distinctions individuals make, whereas explicit gender
stereotypes refer to individuals purposeful and controllable dis-
tinguishment about what is typical of and appropriate for males
and females (Fazio & Olson, 2003). Implicit and explicit gender
stereotypes are only weakly correlated, in part because of explicit
measures' sensitivity to social desirability or lack of insight in one's
own stereotypes (Fazio & Olson, 2003). Teachers can report that
they do not distinguish between male and female students,
whereas observations or implicit measures can prove otherwise
(Jones & Myhill, 2004; Nürnberger, Nerb, Schmitz, Keller, &
Sütterlin, 2016). In a country such as the Netherlands, in which
gender equality is considered important, implicit measures of
gender stereotypes could provide a better understanding of the
(implicit) role of gender stereotypes in teachers' grading behaviors.
In this study, we distinguish between implicit gender-typed
associations and implicit gender-typed expectations. Both are un-
conscious cognitive processes through which one differentiates
between male and female students. Implicit gender-typed associ-
ations refer to teachers' automatic and unconscious associations
between math and masculinity, and between languages and femi-
ninity. Generally, people tend to associate math with masculinity
and languages with femininity based on the traditional belief that
math is the domain in which men can excel and that languages is
the domain in which women can excel (Nosek, Banaji, &
Greenwald, 2002b). As a result, teachers can have stereotypical
perceptions of male and female students' capabilities in math
which can turn in to different assessments of male and female
students (Riegle-Crumb & Humphries, 2012). Math teachers who
have strong associations between math and masculinity and be-
tween language and femininity make a strong distinction between
males' and females’ capabilities in the subjects, which might result
in a stronger gender grading bias. On the other hand, teachers who
have weak associations between gender and subjects might be
more equal in their grading of male and female students.
In a related vein, implicit gender-typed expectations refer to
teachers' automatic and unconscious expectations regarding male
and female students' talent and effort based on gender stereotypes.
Teachers can expect female students to have less talent in math
than male students (e.g., Espinoza, Are ^as da Luz Fontes, & Arms-
Chavez, 2014; Heller, Finsterwald, & Ziegler, 2001; Riegle-Crumb
& Humphries, 2012), and male students to have less talent in lan-
guages than female students (e.g., Schmenk, 2004; Siegle & Reis,
1998). On the other hand, based on negative stereotypes
regarding male students’ interest, motivation, and achievements in
schools (Brown & Stone, 2016), teachers can expect less effort from
3
Teaching and Teacher Education 118 (2022) 103826
male students than from female students (Heyder & Kessels, 2017;
Jones & Myhill, 2004; Siegle & Reis, 1998). Again, teachers who
differ in their expectations of male and female students might
display gender bias in their grading.
Indeed, literature on gender bias in job applications shows that
negative expectations can lead to stricter assessments and positive
expectations can lead to more lenient assessments (Powell, 1986).
These positive or negative expectations can be informed by gender
stereotypes. As a result, female applicants have less chance of being
assessed positively by recruiters for jobs that are perceived as
typically masculine (e.g., management positions), and male appli-
cants have less chance of being assessed positively by recruiters for
jobs that are perceived as typically feminine (e.g., communication
positions) (Cole, Field & Giles, 2004). Following this logic, teachers
who have negative expectations regarding male or female students'
talent or effort in a subject, might assess these students more
strictly, and vice versa. At the same time, teachers with more equal
expectations regarding male and female students’ talent and effort,
might assess male and female students more equally.
In sum, we expect that the stronger a teacher perceives male
and female students as different in terms of their suitability, talent
and effort in math or languages, the stronger the teacher shows
differential grading of male and female students (Dasgupta, 2004;
Gansen, 2019). Vice versa, we expect teachers with weaker gender-
typed associations and gender-typed expectations regarding stu-
dents’ talent and effort to have weaker gender grading bias.
H3. Weaker implicit gender-typed associations and weaker
gender-typed expectations regarding students' talent and effort are
related to weaker gender grading bias.
1.3. Teachers’ gender, gender stereotypes, and gender grading bias
Based on the literature discussed thus far it is likely that implicit
gender-typed associations and gender-typed expectations at least
partly mediate the relation between gender of the teacher and
gender grading bias. That is, the implicit associations between
gender and math and languages may be weaker among female
math teachers and male language teachers compared to male math
teachers and female language teachers (Martin & Dinella, 2012;
Smeding, 2012). Consequently, these weaker implicit gender-typed
associations help shape biased expectations regarding male and
female students' talent and effort in the subjects (Denessen,
Hornstra, Van den Bergh, & Bijlstra, 2022; Muntoni & Retelsdorf,
2018; Van den Bergh, Denessen, Hornstra, Voeten, & Holland,
2010). Teachers' expectations are shaped by a number of co-
occurring factors including teachers' implicit associations
(Denessen et al., 2022). Teachers’ expectations may in turn be
related to gender grading bias. The proposed mediation model is
depicted in Fig. 1.
H4. The association between teachers' gender and gender grading
bias is mediated by teachers' implicit gender-typed associations
and implicit gender-typed expectations.
Female math teachers may have weaker gender-typed associa-
tions, that are in turn related to weaker gender-typed expectations
regarding students' talent and effort, and weaker gender grading
bias than male math teachers. Male Dutch language teachers may
have weaker gender-typed associations, that are in turn related to
weaker gender-typed expectations regarding students’ talent and
effort, and weaker gender grading bias than female Dutch language
teachers.
Our hypotheses concentrated on teachers' gender and implicit
gender stereotypes in relation to gender grading bias without hy-
pothesizing the direction of effects (i.e., effects on grading bias
L. Doornkamp, L.D. Van der Pol, S. Groeneveld et al.
Fig. 1. The indirect effect of teachers' gender on gender grading bias.
against male or female students). Because the examination of
psychological mechanisms that could underly teachers' gender bias
in grading is still largely uncharted territory, we decided to explore
the full spectrum of teachers’ gender stereotypes and gender
grading bias, without excluding participants with contra-
stereotypical associations, expectations, or grading biases a priori.
We thus focused on the extent to which the distinction between
male and female students based on gender stereotypes can result in
differential grading of male and female students. However, our data
allows us to explore the direction of the (mediating) effects of
implicit gender stereotypes both on grading biases against male
and female students in Dutch language and math. These additional
explorations will provide a complete picture of the role of implicit
gender stereotypes in gender grading bias.
2. Methods
This study's field experiment is part of the longitudinal research
project ‘Girls in Science’ that examines gender socialization in the
family and school context in the Netherlands from 2017 until 2022
(ERC funded). The experimental data for this study were collected
between February 2020 and August 2021 among teachers in
training in the subjects Dutch language and math in secondary
education. For the sake of legibility, in the following sections we use
the term ‘teacher/teachers’ to refer to the participating teachers in
training in our study.
2.1. Sample
The teachers who participated in this study were recruited
through their trainers. Trainers from all the 18 institutions that
provide regular teacher training programs for secondary education
in Dutch language and math in the Netherlands were invited to
cooperate by e-mail (trainers from online institutions were not
invited). Thirty-two trainers from 15 institutions agreed to coop-
erate, the others did not participate because of time issues or they
did not respond to the invitation. The trainers either asked their
students (the teachers) to participate in the research or facilitated
the researcher to ask the students (teachers) to participate.
Based on the calls, 159 Dutch language teachers and 268 math
teachers signed up to participate in the research. In all, 28 cases for
Dutch language and 41 cases for math were regarded as missing
data because their evaluations got lost in the Dutch mail system (2
Dutch language, 5 math), they did not send back their evaluations
(24 Dutch language teachers, 26 math), they did not complete the
evaluation (1 Dutch language, 5 math), or they did not complete the
questionnaire (1 Dutch language, 5 math). Only the participants
with scores on all relevant variables were included in the study,
resulting in N ¼ 131 for Dutch language (mean age ¼ 33.73,
SD ¼ 12.098, 76% female), and N ¼ 227 for math (mean age ¼ 29.44,
SD ¼ 12.027, 56% female). Participants that were excluded from the
analyses did not deviate on background characteristics from the
participants that we included.
The distribution of male and female teachers across math and
Dutch language is skewed, but similar to the gender distribution in
Dutch secondary schools. Dutch language teachers in the
Netherlands are disproportionately female (72%). For math, the
distribution is slightly more balanced with 58% male and 42%
4
Teaching and Teacher Education 118 (2022) 103826
female teachers (numbers based on open educational data 2018:
DUO, 2020).
Both the Dutch language and math teachers performed intern-
ships at secondary schools for 10 h (Dutch language SD ¼ 8.53,
math SD ¼ 7.49) a week. For Dutch language, 27% of the participants
received their training on a university (first degree) whereas 73%
received their training on a university for applied sciences (second
degree). Furthermore, 41% of the Dutch language teachers had
experience with teaching. For math, 15% received their training at
university, whereas the other 85% received it at a university for
applied sciences. In math, 37% of the teachers in training had
experience with teaching.
2.2. study design
The participating teachers each (1) graded a set of 24 tests made
by last-year students in higher secondary education, (2) assessed
each test on invested talent and effort, and (3) filled in an online
questionnaire. Data collection took place at home and participation
required approximately 2.5 h. In return for their participation,
teachers received individual feedback on their grading practices,
the results of the study, and a small gift.
The student names above the 24 tests were manipulated.
Originally, 10 female students and 14 male students completed the
Dutch language tests and 14 female students and 10 male students
completed the math tests. The 24 tests were divided into two
groups with the same average scores; one group was given names
that are generally considered as male, whereas the other group was
given names that are generally considered as female. We tried to
maximize the ecological validity of our experiment within the
limitations of its practical feasibility. To create conditions that are
transferable to the real-life exam evaluation situation, we designed
a grading experiment using real school exams with real students’
answers. The study design is depicted in Fig. 2.
The Dutch language and math school exams and students’ an-
swers to the exams that were used in the experiment were
retrieved from a Dutch language and a math teacher from different
Dutch secondary schools in the western region of the Netherlands.
Both exams were conducted in 2019 by students in the final year of
their secondary education.
We selected six open-answer questions from both school exams
(Dutch language and math) to avoid asking too much time from the
participants. We used the following two criteria in the selection of
these questions: (1) there had to be ample room for participants’
own interpretation in their assessment of the given answers to the
question, and (2) there was sufficient between-student range in the
points given for answers to a question by the original assessor to
ensure that the quality of the answers to the question differed.
From the groups of students who originally took the Dutch
language exam (n ¼ 39) and the math exam (n ¼ 45), 24 students
for each subject were selected to mimic real-life situations in which
teachers must evaluate the tests of one class with approximately 25
students. The selected students’ answers (including mistakes,
misspellings, and notes) were copied into typed text and printed to
eliminate any reference to the gender of the original student
through for example handwriting.
To ensure that we could form two groups of tests (one group
with names that are generally considered as male and one group
L. Doornkamp, L.D. Van der Pol, S. Groeneveld et al.
Fig. 2. Study design.
with names that are generally considered as female) of equal
quality in terms of answers given to the selected questions, we
recruited 10 experienced teachers for each subject to blindly grade
the 24 tests (i.e., without student names). Based on these blind
evaluations, two groups were created with 11 tests each.
We used common Dutch female names (8), Dutch male names
(8), Turkish female names (4), and Turkish male names (4) that
were randomly distributed over the groups of tests (for an overview
of the names see Appendix A). To reflect the Dutch multicultural
society, we used Turkish names in addition to Dutch names as
people with a Turkish migration background are, after the native
Dutch, the largest ethnic-cultural group in Dutch society (CBS,
2022). To control for an unforeseen effect of differences in stu-
dents’ answers we created two versions. The tests with male names
in the first version had female names in the second version, and the
tests with female names in the first version had male names in the
second version. The versions were randomly distributed across the
participants. Furthermore, to control for grading order and grading
fatigue, the order of tests was randomly distributed and different
for the two versions. Additionally, the first test in the set alternated
within each version.
In addition to the evaluation of the tests, participants were
asked to estimate students’ talent and effort based on each test. The
24 tests, including these questions, were presented in a booklet. A
short introduction with general information about the experiment
was added, but nothing was said about the gender grading bias
component. To prevent participants from being primed, partici-
pants were told that the study was on the topic of grading accuracy
in general. After completing their participation in the study,
teachers in training were fully informed about the goal of the study.
The booklet also included an answer model provided by the
original assessors, with a brief description of the main elements of a
correct answer for each question. The maximum number of points
per question for a good answer (which was similar to the original
assessment) was given to the participants, but not further specified.
The maximum total number of points for the Dutch language exam
was 10 whereas for math it was 26.
During the evaluations of the tests, participants were asked to
write the name of the student on a provided form. This was done to
ensure that participants read the name of the student and thus
became aware of the student's gender (the manipulation). Ethical
approval for this study was provided by the Ethics Committee of the
Faculty of Governance and Global Affairs (University Leiden).
2.3. Measures
2.3.1. Gender grading bias
Gender grading bias refers to bias in evaluations against either
male or female students. The mean scores for the tests with a male
5
Teaching and Teacher Education 118 (2022) 103826
name (11) and the tests with a female name (11) were calculated
based on the total number of points rewarded for the answers to
the six questions. Then, gender grading bias was calculated by
subtracting the mean male test score from the mean female test
score. As a result, gender grading bias is a continuous variable
including negative and positive scores. The negative scores indicate
a gender grading bias against females, whereas positive scores
indicate a gender grading bias against males.
2.3.2. Implicit gender-typed expectations
The gender stereotypical expectations of the teachers were
measured for two aspects: talent for the subject and invested effort.
After the teacher evaluated a single test, the teacher was asked to
estimate the (fictious) student's talent and effort on a scale from 1
to 10 (i.e., on a scale from 1 to 10 to what extent do you think this
student has talent for the subject/invested time in preparing the
test). We did not put any explicit emphasis on gender of the
student.
Similar to the gender grading bias measure, mean scores were
calculated for the tests with male names (11) and the tests with
female names (11). Then, gender bias in expectations was calcu-
lated by subtracting the mean score for the male tests from the
mean score on the female tests. Again, negative scores indicate a
bias against females (males have more talent/put more effort than
females), whereas positive scores indicate a bias against males
(females have more talent/put more effort than males).
2.3.3. Implicit gender-typed associations
The implicit gender-typed associations of the teachers were
assessed with the arts-science Implicit Association Test (IAT). The
IAT was built in LimeSurvey, an online questionnaire tool, and was
based on the task on the Harvard project Implicit demonstration
website and the Nosek, Banaji, and Greenwald (2002a) paper. The
IAT measures participants' implicit association between female and
male attributes and school related concepts ‘languages’ and ‘sci-
ence’. During the IAT, participants were requested to sort words
into groups by pressing keys. In congruent blocks, female names
(e.g., ‘Julia’) and stereotypically feminine words (e.g., ‘Dutch lan-
guage’) needed to be sorted under the ‘Female & Language’ cate-
gory and male names (e.g., ‘Daan’) and stereotypically masculine
words (e.g., ‘math’) needed to be sorted under the ‘Male & Science’
category. In incongruent blocks female names and stereotypically
masculine words needed to be sorted under the ‘Female & Science’
category and male names and stereotypically feminine words
needed to be sorted under the ‘Male & Languages’ category. The
reaction time and accuracy scores were recorded by LimeSurvey.
Each participant's level of implicit gender-typed associations was
determined by calculating differences in reaction time and accuracy
in congruent and incongruent blocks resulting in a d-score (using
L. Doornkamp, L.D. Van der Pol, S. Groeneveld et al.
scoring algorithm by Greenwald, Nosek, & Banaji, 2003). A positive
d-score represents stronger implicit gender-typed associations
(e.g., math is for males, Dutch language for females). Negative d-
scores represent counter-gender-typed associations (e.g., math is
for females, Dutch language for males). To reduce possible order
effects of the presentation of congruent and incongruent blocks,
two versions of the IAT were used, one in which the congruent
block was administrated first and one in which the incongruent
block was administrated first (Nosek, Greenwald, & Banaji, 2005).
2.4. Data analyses
We used SPSS Statistics version 27 for data inspection and
further analyses. All variables were inspected for possible outliers
that were defined as values more than 3.29 SD under or above the
mean (Tabachnick & Fidell, 2012). Two outliers were identified and
winsorized by giving them the most extreme not outlying value
(Tabachnick & Fidell, 2012). All continuous variables were normally
distributed.
To examine whether teachers' grading bias is negatively asso-
ciated with male students we plotted the distribution of the scores
of the gender grading bias variable in histograms for Dutch lan-
guage and math. Additional one sample t-tests were performed to
test whether the mean scores were significantly different from zero
(hypothesis 1). Differences between male and female teachers’
gender grading bias in Dutch language and math were examined
using independent t-tests (hypothesis 2).
To test hypotheses 3 and 4, the serial multiple mediation model
(Model 6) in PROCESS was run to incorporate bootstrapping tech-
niques for estimating indirect effects (Hayes, 2018). Using this
model we estimated the effects of implicit gender-typed associa-
tions and expectations on gender grading bias controlled for
teachers' gender (hypothesis 3), as well as the indirect effect of
teachers' gender on gender grading bias through implicit gender-
typed associations and expectations (hypothesis 4). We ran the
model for expectations regarding students’ talent and effort sepa-
rately. Indirect effects are significant when the confidence intervals
do not include zero.
The main analyses provided insight in the strength of the as-
sociations between the main variables, but lacked information
about the direction of the effects of implicit gender stereotypes on
grading biases against either male or female students. To explore
these directions, scatterplots with regression lines were produced
and additional post hoc analyses were performed with separate
outcome variables for grading bias against male students and
grading bias against female students (results are presented in
appendix B). The gender grading bias variable in the main analyses
included scores from 3 to þ3 with negative scores representing
grading bias against females and positive scores representing
grading bias against male students. In the post hoc analyses, the
gender grading bias variable was separated for a grading bias
against male students and a grading bias against female students. In
the variable for grading bias against male students, we gave
teachers with biases against female students score 0. In the variable
for grading bias against female students, we gave teachers with
biases against male students score 0. This corresponds with the
method of Liben and Bigler (2002) to infer stereotypical attitudes
towards others, rather than contra-stereotypical attitudes, using
questionnaire data. A higher score on the variables in the post hoc
analyses indicated a stronger grading bias against either male or
female students.
To test the robustness of our results we conducted the analyses
with the control variables age, experience, and test version.
6
Teaching and Teacher Education 118 (2022) 103826
3. Results
3.1. Gender grading bias on the aggregate level
Table 1 presents the means, standard deviations and range of
gender grading bias in Dutch language and math. The distributions
of the scores on gender grading bias are depicted in Fig. 3. Negative
scores indicate a grading bias against female students, positive
scores indicate a grading bias against male students. In Dutch
language, the mean gender bias in grading is 0.023 (SD ¼ 0.533),
indicating that Dutch language teachers on average did not differ-
entiate in their grading based on students’ gender (t (130) ¼ 0.497,
p ¼ 0.620, d ¼ 0.043). In math, the average score on the gender
grading bias variable is 0.019 (SD ¼ 0.828), indicating that math
teachers did not differentiate in their grading of male and female
students either (t (226) ¼ 0.338, p ¼ 0.736, d ¼ 0.022. Thus, hy-
pothesis 1 should be rejected.
Nevertheless, Table 1 and Fig. 3 show that there is ample vari-
ation in gender grading biases on the individual level. In Dutch
language, teachers’ grading varies from a bias against female stu-
dents, rewarding male students up to 1.14 point higher than female
students, to a bias against male students, rewarding female stu-
dents up to 1.64 points higher than female students (on a scale from
1 to 10). In math, in the most extreme cases math teachers graded
male students up to 2.36 points higher than female students, or
graded female students up to 2.64 points higher than male students
(on a scale from 1 to 26).
3.2. Teachers’ gender and gender grading bias
Next, we tested whether the variation in teachers' gender
grading bias is associated with teachers' gender. Table 1 reveals that
teachers' gender grading bias is not correlated with teachers’
gender. Independent sample t-tests confirm that gender grading
bias among male Dutch language teachers (M ¼ 0.047, SD ¼ 0.519)
is not significantly different from the gender grading bias among
female Dutch language teachers (M ¼ 0.016, SD ¼ 0.540), t
(129) ¼ 0.283, p ¼ 0.616, d ¼ 0.058. Also for math, gender grading
bias among male teachers (M ¼ 0.007, SD ¼ 0.838) is not signifi-
cantly different from gender grading bias among female teachers
(M ¼ 0.027, SD ¼ 0.824), t (225) ¼ 0.180, p ¼ 0.858, d ¼ 0.024.
Therefore, hypothesis 2 should be rejected.
3.3. Teachers’ implicit gender stereotypes and gender grading bias
Table 1 presents the mean scores and correlations of teachers'
implicit gender-typed associations and implicit gender-typed ex-
pectations. On average, teachers in both subjects tend to associate
math with masculinity and languages with femininity (Dutch lan-
guage t (130) ¼ 13.158, p ¼ 0.000, d ¼ 1.150, math t (226) ¼ 13.487,
p ¼ 0.000, d ¼ 0.895. Further, both Dutch language and math
teachers expect female students to invest more effort in preparing
an exam than male students (Dutch language t (130) ¼ 2.622,
p ¼ 0.010, d ¼ 0.229, math t (226) ¼ 8.279, p ¼ 0.000, d ¼ 0.549).
Additionally, math teachers expect male students to have more
talent in math than female students (t (226) ¼ 4.043, p ¼ 0.000,
d ¼ 0.268). Table 1 shows that teachers' gender-typed associa-
tions are not correlated with gender grading bias in either Dutch
language or math. However, teachers' gender-typed expectations
were positively correlated with gender grading bias. In Dutch lan-
guage, both gender-typed expectations regarding students' talent
and effort were positively associated with teachers' gender bias in
grading. In math, only teachers' expectations regarding students’
talent were positively associated with gender grading bias.
Models 3 and 6 in Table 2 present the results of regressing
L. Doornkamp, L.D. Van der Pol, S. Groeneveld et al. Teaching and Teacher Education 118 (2022) 103826

Table 1
Mean Scores and Pearson Correlations on Central Variables of the Study.

Mean (SD) Min, Max Pearson Correlations

1 2 3 4

Dutch 1 Gender Grading Bias 0.023 (0.533) 1.14, 1.64

Language 2 Gender Teacher 0.760 (0.427) 0, 1 0.025
N ¼ 131 3 Gender-typed Associations 0.488 (0.424) 0.51, 1.35 0.069 0.512**
4 Expectations Talent 0.038 (0.490) 1.39, 1.55 0.727** 0.050 0.008
5 Expectations Effort 0.114 (0.499) 1.36, 1.25 0.410** 0.119 0.004 0.404**

Math 1 Gender Grading Bias 0.019 (0.828) 2.36, 2.64

N ¼ 227 2 Gender Teacher 0.560 (0.497) 0, 1 0.012
3 Gender-typed Associations 0.386 (0.431) 0.74, 1.33 0.031 0.311**
4 Expectations Talent 0.129 (0.480) 1.64, 1.09 0.442** 0.024 0.119
5 Expectations Effort 0.284 (0.517) 1.27, 1.82 0.116 0.004 0.122 0.042

Note: Gender, 0 ¼ male, 1 ¼ female, Expectations Talent ¼ Implicit gender-typed expectations regarding students' talent, Expectations Effort ¼ Implicit gender-typed ex-
pectations regarding students' effort. **p < 0.01, *p < 0.05.

Fig. 3. Histograms gender grading bias.

Table 2
Double Mediation of Implicit Gender-typed Associations and Expectations regarding Talent (top-half) and Effort (bottom-half) in the Relation Between Teachers’ Gender and
Gender Grading Bias.

Dutch Language (N ¼ 131) Math (N ¼ 227)

(1) IA (2) IE (3) GGB (4) IA (5) IE (6) GGB

Constant 0.099 (0.066) 0.076 (0.089) 0.027 (0.067) 0.539 (0.041) 0.032 (0.064) 0.063 (0.100)
Gender Teacher 0.509** (0.075) 0.084 (0.118) 0.036 (0.088) 0.270** (0.055) 0.065 (0.067) 0.055 (0.106)
Implicit Associations 0.053 (0.119) 0.098 (0.089) 0.155* (0.077) 0.063 (0.122)
Expectations Talent 0.789** (0.066) 0.772** (0.105)
R2 0.262** 0.004 0.534** 0.097** 0.018 0.197**

Constant 0.099 (0.066) 0.210 (0.090) 0.059 (0.091) 0.539 (0.041) 0.194 (0.069) 0.001 (0.112)
Gender Teacher 0.509** (0.075) 0.192 (0.119) 0.018 (0.119) 0.270** (0.055) 0.048 (0.073) 0.005 (0.117)
Implicit Associations 0.104 (0.120) 0.094 (0.118) 0.163 (0.084) 0.089 (0.135)
Expectations Effort 0.436** (0.087) 0.194 (0.108)
R2 0.262** 0.020 0.173** 0.097** 0.017 0.015

Note: Bootstrap estimates derived from 10 000 bootstrap samples along with 95% CIs are reported. Standard Deviations in parentheses. IA ¼ Implicit Gender-typed Asso-
ciations, IE ¼ Implicit Gender-typed Expectations, GGB ¼ Gender Grading Bias, **p < 0.01, *p < 0.05.

teachers' implicit gender-typed associations and implicit gender-
typed expectations regarding students' talent (top half of Table 2)
and effort (bottom half of Table 2) on gender grading bias
controlled for teachers' gender (hypothesis 3). For Dutch language,
Model 3 confirms that teachers' implicit gender-typed associations
are not related to gender grading bias. The extent to which teachers
associate math with masculinity and languages with femininity is
not related to gender bias in teachers' grading. This Model also
confirms that teachers' gender-typed expectations regarding
students' talent and effort are positively associated with gender
grading bias. Fig. 4 indicates that teachers who expect male stu-
dents to have more talent in Dutch language than female students,
give male students more points for their answers than female
students, despite the equal quality of the tests. Teachers who expect
female students to have more talent in Dutch language than male
students, appear to give female students more points for their an-
swers than male students. The same is true for the relation between
teachers' gender-typed expectations regarding students’ effort and

7
L. Doornkamp, L.D. Van der Pol, S. Groeneveld et al. Teaching and Teacher Education 118 (2022) 103826

Fig. 4. Plots of the relation between gender-typed expectations and gender grading bias Note: negative scores indicate biases against female students, positive scores indicate biases
against male students.

gender grading bias. Post-hoc analyses confirm this image (see gender-typed expectations regarding students’ invested effort is
appendix B). mixed.
For math, Model 6 in Table 2 shows a similar pattern. Teachers'
implicit gender-typed associations are not related to gender
3.4. The indirect relation between gender and gender grading bias
grading bias, neither are teachers' gender-typed expectations
regarding students' effort. However, teachers' gender-typed ex-
Table 1 reveals that teachers' gender is correlated with teachers’
pectations regarding students' talent is positively associated with
gender-typed associations in Dutch language and math. In Dutch
gender grading bias. Also for math, Fig. 4 indicates that the positive
language, gender and gender-typed associations are positively
relation between teachers' expectations and gender grading bias
correlated, indicating that male Dutch language teachers have
works both ways: teachers who expect male students to have more
weaker gender-typed associations than female Dutch language
talent in math compared to female students, give male students
teachers. In math, gender and gender-typed associations are
more points for their answers compared to female students.
negatively correlated, indicating that female math teachers have
Teachers who expect female students to have more talent in math
weaker gender-typed associations than male math teachers.
than male students, give female students more point for their an-
Despite the absence of a direct effect of teachers' gender on their
swers compared to male students. Post-hoc analyses confirm this
gender grading bias, we estimated the indirect effect of teachers'
pattern regarding the association between teachers' expectations
gender on gender grading bias because a correlation between X and
of students' talent and gender grading bias. Furthermore, the post-
Y is not necessarily needed as a precondition for a mediation effect
hoc analyses show that teachers' gender-typed expectations
(Hayes, 2018). Table 2 presents the findings. Models 1 and 4 include
regarding students’ effort are significantly positively associated
the findings of regressing teachers' gender on their implicit gender-
with a grading bias against female students, but not with a grading
typed associations. Models 2 and 5 present the findings of
bias against male students. Teachers who expect male students to
regressing both teachers' gender and implicit gender-typed asso-
put more effort in preparing for an exam than female students give
ciations on their gender-typed expectations regarding students’
male students more points for their answers than female students,
talent (top half of Table 2) and effort (bottom half of Table 2). As
whereas teachers who expect female students to put more effort in
mentioned above, Models 3 and 6 show the results of regressing
preparing for an exam than male students do not give female stu-
gender, implicit gender-typed associations and gender-typed ex-
dents more points for their answers than male students (see
pectations on gender grading bias. The mediation effects of implicit
appendix B).
gender-typed associations and gender-typed expectations, as well
In all, hypothesis 3 can only partly be accepted. Implicit gender-
as the double mediation effect and the total effect of the model are
typed associations are not directly associated with gender grading
presented in Table 3.
bias. However, teachers' gender-typed expectations regarding stu-
For Dutch language, the findings are presented in the first three
dents' talent is related to gender grading bias. Evidence for a role of
Models. Model 1 confirms that implicit gender-typed associations
8
L. Doornkamp, L.D. Van der Pol, S. Groeneveld et al. Teaching and Teacher Education 118 (2022) 103826

Table 3
Indirect effects of gender on gender grading bias.

Dutch Language Math

Effect Size (SE) LLCI ULCI Effect Size (SE) LLCI ULCI

Total Mediated Effect 0.004 (0.098) 0.192 0.192 0.035 (0.055) 0.145 0.073
Mediation through IA 0.050 (0.050) 0.051 0.146 0.017 (0.032) 0.084 0.046
Mediation through IE 0.066 (0.086) 0.239 0.098 0.050 (0.053) 0.155 0.053
Double mediation 0.021 (0.042) 0.059 0.110 0.032 (0.019) 0.000 0.073

Total Mediated Effect 0.013 (0.087) 0.187 0.155 0.025 (0.042) 0.055 0.111
Mediation through IA 0.048 (0.068) 0.090 0.182 0.024 (0.039) 0.052 0.107
Mediation through IE 0.084 (0.050) 0.186 0.011 0.009 (0.017) 0.022 0.051
Double mediation 0.023 (0.032) 0.032 0.095 0.009 (0.007) 0.026 0.002

Note: Total Mediated Effect is the sum of the indirect effects of X > Y. Size of bootstrap sample for calculations of the indirect effect ¼ 10 000. LLCI ¼ bootstrapped lower level
confidence interval; ULCI ¼ bootstrapped upper level confidence interval, 95% confidence interval, Effect size is significant when LLCI and ULCI do not include 0.

are stronger among female Dutch language teachers compared to related to having a stronger grading bias against female students
male Dutch language teachers. As expected, male Dutch language (see appendix B). In all, hypothesis 4 should be accepted in the
teachers have less stereotypic associations between gender and subject math, but rejected in the subject Dutch language.
math and languages. Model 2 shows that both gender and implicit Finally, the pattern of results for the model was similar when
gender-typed associations are not related to gender-typed expec- controlling for teachers' age, experience (Lavy, 2008), and the
tations regarding students' talent and effort. Contrary to what was version of the tests that were graded by the teachers, though it
expected, strong implicit gender-typed associations are not related should be noted that for math the significance level of the effect of
to making a stronger distinction between male and female students implicit gender-typed associations on gender-typed expectations
regarding their talent and effort in Dutch language. As described regarding students’ talent and the double mediation effect were no
above, Model 3 shows that teachers’ gender-typed expectations are longer significant on the 0.05 level, but on the 0.1 level.
related to gender grading bias.
For math, the findings are presented in the last three Models.
4. Conclusion and discussion
Model 4 confirms that implicit gender-typed associations are
stronger among male math teachers compared to their female
By means of a field experiment among teachers in training in the
counterparts. As expected, female math teachers have milder
subjects Dutch language and math in secondary education in the
gender-typed associations than male math teachers. Model 5 re-
Netherlands, this study tested the influence of students' gender on
veals that teachers' gender is not associated with gender-typed
teachers' grading and examined the role of teachers' gender and
associations regarding students' talent and effort in math. Teach-
implicit gender stereotypes in gender grading bias. We found that
ers' implicit gender-typed associations are negatively related to
teachers' grading, on average, was not gender biased. However, we
their gender-typed expectations regarding students' talent, but not
did find that individual variation in gender grading bias was asso-
related to teachers' expectations regarding students' effort. As
ciated with teachers' implicit expectations of male and female
described above, Model 3 shows that only teachers' gender-typed
students' talent and effort. Further, we found an indirect effect of
expectations regarding students’ talent are positively related to
teachers' gender on gender grading bias in math via teachers’ im-
gender grading bias.
plicit gender-typed associations and expectations.
Table 3 summarizes the effects of teachers' gender on gender
Contrary to what we expected (e.g., Protivínský & Münich,
grading bias, through implicit gender-typed associations and ex-
2018), in our field experiment we did not find an association be-
pectations. For Dutch language, no indirect effects were found for
tween perceived male gender of students and teachers' grading.
implicit gender-typed associations and gender-typed expectations
Instead, we found that teachers' grading, on average, did not differ
(or both). For math, implicit gender-typed associations and ex-
between male and female students. This finding is similar to the
pectations could not explain differences in male and female
few studies that also experimentally tested the influence of stu-
teachers' gender grading bias on their own. However, a positive
dents' gender on teachers' grading (Hanna & Linden, 2012;
double mediation effect was found: differences in male and female
Hinnerich et al., 2011), suggesting that students' gender and
teachers' gender grading bias can be explained through differences
teachers' grading might not be causally related. Previous studies
in male and female teachers' implicit gender-typed associations
that found grading bias against male students, all used comparative
and expectations regarding students' talent. Post-hoc analyses
designs to examine gender grading bias, i.e., comparing scores on
reveal that teachers' implicit gender-typed associations and ex-
tests that were blindly graded to scores on tests that were graded
pectations regarding students' talent can explain differences in
by a teacher (e.g., Protivínský & Münich, 2018). The observed bias
male and female math teachers' grading biases against male stu-
against male students in teachers' grading in these studies might
dents and female students. Both for the grading bias against male
not be a direct result of students' gender but rather a reflection of
students and the grading bias against female students applies that
teachers' beliefs about male and female students. Indeed, one of
female math teachers have weaker associations between math and
these previous studies found that an effect of grading bias against
masculinity, that is related to expecting approximately the same of
male students disappeared when controlling for teachers' reports
male and female students in terms of talent, which is related to
on students' attitudes towards learning (Cornwell et al., 2013).
having weaker (closer to zero) gender grading bias. Furthermore,
Although Cornwell et al. (2013) concluded that differences in male
the post-hoc analyses show that implicit gender-typed associations
and female students' behavior could fully explain gender bias in
and expectations regarding students’ effort can explain differences
teachers' grading, based on the results in our study we argue that it
in male and female teachers grading biases against female students.
could have been a gender bias in teachers' perception of male and
Female math teachers have weaker gender-typed associations that
female students’ behavior that explained gender grading bias.
lead to expecting male students to put less effort in math which is
Although we did not find a gender bias in teachers' grading on
9
L. Doornkamp, L.D. Van der Pol, S. Groeneveld et al.
average, we did find ample variation in gender grading bias be-
tween teachers. Our study showed that the variation in gender
grading bias was associated with variation in teachers' expectations
of male and female students’ talent and effort despite the equal
quality of the tests of male and female students. Teachers with
more equal expectations of male and female students displayed less
gender bias in grading. Teachers who differed in their expectations
of male and female students displayed more grading bias against
both male and female students, despite the equal quality of their
tests. Our finding confirmed previous research that demonstrated
that the extent to which a person distinguishes between males and
females based on generalized preconceptions is related to their
discriminative behaviors against males and females (e.g., Burgess &
Borgida, 1999; Dasgupta, 2004), and that different expectations can
lead to different assessments (Muntoni & Retelsdorf, 2018; Powell,
1986).
Additionally, to get a complete picture of the role of implicit
gender-typed expectations in gender grading bias our study
explored the full spectrum of gender-typed expectations including
stereotypical and contra-stereotypical expectations. Based on pre-
vious research we expected gender stereotypes to differ mainly in
strength, and to lesser extent in direction (e.g., Lenton et al., 2009).
However, the post-hoc analyses showed that teachers' implicit
gender-typed expectations differed in both strength and direction,
and that both stereotypical and contra-stereotypical expectations
were related to discriminative behaviors. Dutch language and math
teachers' grading biases against male or female students were in
line with the content of their biases in expectations. Despite the
equal quality of the tests, in both subjects, teachers who expected
male students to have more talent, gave male students more points
for their answers. Teachers who expected female students to have
more talent, gave female students more points for their answers.
The same was true for Dutch language teachers' expectations
regarding students' effort: expecting more effort from male stu-
dents was related to rewarding male students more points for their
answers, and expecting more effort from female students was
related to rewarding female students more points for their answers.
The explorations support previous research in the family context
that found that fathers' stereotypical and contra-stereotypical ex-
pectations of males and females are related to opposite differential
treatments of males and females (Endendijk et al., 2013). It should
be noted that, in math, only teachers' contra-stereotypical expec-
tations of students' effort were associated with gender biases
against female students. Math teachers who had the contra-
stereotypical expectation that male students put more effort in
the preparation of a test, gave male students more points for their
answers, whereas math teachers who expected female students to
put more effort, did not gave female students more points for their
answers. Putting effort is gender stereotyped as something female
students generally do, and male students generally do not do (e.g.,
Heyder and Kessels, 2007). Therefore, expecting students to have
put effort could be praised and result in positive effects in teachers’
grading for male students but not for female students. This finding
supports literature on double standards for men and women in,
among others, parenting roles (Deutsch & Saxon, 1998), on the
work floor (Lyness & Heilman, 2006), and in engaging in sexual
behaviors (Endendijk, Van Baar, & Dekovi c, 2020; Sagebin Bordini &
Sperb, 2013).
Furthermore, similar to previous research (Martin & Dinella,
2012; Smeding, 2012), our results showed that female teachers in
a stereotypically masculine subject (math) and male teachers in a
10
Teaching and Teacher Education 118 (2022) 103826
stereotypically feminine subject (languages) had weaker associa-
tions between math and masculinity and between language and
femininity than male math teachers and female Dutch language
teachers. Our finding supports the assumption that an individual's
gender stereotypes are flexible and that these gender stereotypes
can become less strict by experiencing or observing rejection of the
gender stereotype (Bussey & Bandura, 1999; Charlesworth & Banaji,
2021; Groeneveld, Van der Pol, De Vries, & Mesman, 2021). Though
it should be noted that some scholars have argued that gender
stereotypes are resistant to change (Fiske, 2017; Haines, Deaux, &
Lofaro, 2016). Additionally, previous studies showed that implicit
gender stereotypes were similar among males and females (Banaji
& Greenwald, 1995; Rudman & Kilianski, 2000), or somewhat
stronger among females than males (Endendijk et al., 2013; Nosek
et al., 2002a). Our study showed that both males and females can
have weaker implicit gender stereotypes and that these seem to be
dependent on someone's own experiences rather than someone's
gender perse.
We found that these weaker implicit gender-typed associations
among female math teachers were related to weaker gender-typed
expectations of students and that this mechanism could explain
weaker gender grading biases among female math teachers than
male math teachers. Explorations of the direction of the mediating
effect showed that, in math, female teachers' weaker gender ste-
reotypical associations (math is for men and women, languages is
for men and women) were related to weaker gender-stereotypical
expectations of students' talent (males and females are talented in
math). This finding supports previous findings that teachers' ste-
reotypes and expectations are positively related (Muntoni &
Retelsdorf, 2018). Consequently, weaker (more equal) gender-
typed expectations regarding students' talent resulted in a
weaker grading bias against both males and females. Weaker
gender-typed expectations regarding students' effort could only
explain differences between male and female math teachers’
grading bias against female students, but it should be noted that
this effect size was fairly small.
To our knowledge, our study is the first to find evidence for a
psychological mechanism underlying gender grading bias. How-
ever, we did not find evidence for the mechanism in Dutch lan-
guage teachers. Male Dutch language teachers did have weaker
gender-typed associations than female Dutch language teachers,
but these were not related to weaker gender-typed expectations.
This may be because gender stereotypes regarding students’ talent
in language are not as persuasive as in math. Large government
campaigns were directed at enhancing female participation in
math, but not at enhancing male participation in languages.
Furthermore, many famous Dutch writers are male, providing
convincing contra-gender stereotypical examples in Dutch lan-
guage. It thus seems that gender salience is important for effects of
implicit gender stereotypes in gender grading bias to occur (Keiser,
Wilkins, Meier, & Holland, 2002).
The findings of this study should be interpreted in light of its
limitations. The balancing between the maximalization of this ex-
periment's ecological validity and practical feasibility resulted in
the exclusion of real-life student-teacher interactions that could
play an important role in gender grading bias (Falch & Naper, 2013).
Moreover, the study design resulted in the selection of two
experimental groups of tests that were not of the exact same
quality. The group of tests with male names were of approximately
the same quality as the group of tests with female names based on
the experts' blind grading. However, as the tests in the two groups
L. Doornkamp, L.D. Van der Pol, S. Groeneveld et al.
were not identical in terms of students' answers, differences in the
students' answers might have resulted into differences between
the two groups that affect gender grading bias. Further, although
implicit measures are encouraged to study differential treatment of
male and female students (Denessen et al., 2022), in particular the
IAT is criticized for its validity and reliability in predicting
discriminative behaviors (Blanton et al., 2009). However, for the
current study the Arts-Science version of the IAT was highly rele-
vant because of the theoretical link between the content of this test
and our outcome measure, i.e., gender grading bias in the subjects
math and languages. Despite the limitations of the use of different
and real-student answers and the implicit measures of gender
stereotypes, these components of this study's design contributed in
particular to the participants not being aware of what the research
was about. Therefore, social desirability and conscious behaviors,
which participants often have when they participate in social
research (Wulff & Villadsen, 2019), were probably prevented.
In all, the results in our study question previous research on the
role of students' behaviors in gender grading bias that relied on
teacher reports (Cornwell et al., 2013). Our study encourages future
research to be conscious of gender bias in teacher reports when
relying on teacher reports to explain differences between male and
female students. Furthermore, we recommend further exploration
of the psychological mechanisms underlying gender grading bias
and empirically test the role of gender stereotypes and teachers'
gendered beliefs. Future research could tap into the role of teachers'
explicit gender stereotypes, or the extent to which teachers'
gendered beliefs and gender grading bias are causally related to
unravel the processes that can result in gender biases in teachers’
grading.
Further, our results have implications for teacher training pro-
grams. Teacher training programs might want to contribute to
increasing awareness of teachers (in training) about the role of their
own gender-typed expectations of male and female students in
grading practices. Improving teachers' self-awareness can stimu-
late their quality of teaching (Schussler, Stooksberry, & Bercaw,
2010). We are working on translating the setup of this study's
experiment into an educational tool directed at increasing teachers'
awareness of their own gender (grading) bias and improving
teachers agency in equal treatment of male and female students.
Teachers achieve agency (i.e., actions that are assumed to
contribute to good and meaningful education (Biesta, Priestley, &
Robinson, 2015) when they are able to intentionally choose be-
haviors instead of just behave in a routinized matter without
considering alternatives (Priestley, Biesta, & Robinson, 2015).
Awareness of one's own gender-typed expectations of male and
female students and gender grading bias, might help teachers to
tackle automatic thoughts and behaviors and to treat males and
females more equally, both in interaction with and evaluation of
students. Indeed, previous research demonstrated that awareness
Dutch male names Dutch female names
1 Bastiaan Driessen Sara Wolters
2 Thomas van der Horst Meike Groen
3 Danie€l van der Pol Juliette Kramer
4 Benjamin de Lange Madelief Scholten
5 Pieter Smeets Fien de Ruiter
6 Teun Bakker Emma Hoekstra
7 Lars Willemsen Tess van Veen
8 Florian van Dijk Olivia Maas
11
Teaching and Teacher Education 118 (2022) 103826
of one's own gender bias can contribute to reducing gender bias in
an education setting (Carnes et al., 2015; Devine et al., 2017;
Forscher, Mitamura, Dix, Cox, & Devine, 2017; Girod et al., 2016). It
should be noted that it is crucial that gender equality is valued by
the individual teacher as well as by the (school) collective (Biesta
et al., 2015). This can be a complicating factor as the extent to
which gender inequality is perceived (see for instance Kinias & Kim,
2011), gender equality in education is valued, and ideas about the
routes towards equal treatment can differ between individual
teachers and across cultures (see Cardona Lo  pez, Nordfjell, Gaini, &
Heikkinen, 2018). Besides the importance of discussions between
teachers, teacher trainers, and policy makers ‘to get everyone on
the same page’ (Cardona Lo  pez et al., 2018), we believe that courses
that increase awareness of teachers' own gender biases are critical
as gender bias often happens unconsciously and its consequences
are therefore often overlooked or underestimated. Teacher training
programs could be a suitable place to increase teachers' (in
training) self-awareness of gender stereotypes and agency in equal
treatment of boys and girls (De Boer, Janssen, Dam, & Van Driel,
2019).
To conclude, our study pointed out that rather than a causal
relation between students' gender and teachers grading, the role of
teachers' implicit gendered beliefs about male and female students
should be considered. Not only were teachers' implicit gender-
typed expectations related to gender grading bias, implicit
gender-typed associations and expectations together could explain
differences in male and female math teachers' grading biases.
Students' gender and teachers' gender do not directly translate in
differential grading practices, instead, teachers’ beliefs about males
and females can affect their perception and assessment of male and
female students.
Funding
This work was supported and funded by the European Research
Council (ERC), project Girls in Science (grantnumber 726141). The
funders had no role in the study design, data, analysis and inter-
pretation or preparation of the manuscript.
Acknowledgements
The authors would like to thank Arthur Pormes, Anneke Wurth
and Peter Kop for their assistance in the design of the field exper-
iment. Further, the authors would like to thank Anne Floor Lubbers
for her effort during the experiment's data collection.
Appendix A
Turkish male names Turkish female names
Yusuf Koc Leyla Bulut
Murat Aslan Sevda Avci
Hakan Yilmaz Kubra Demir
Ahmet Cicek Merve Günes
L. Doornkamp, L.D. Van der Pol, S. Groeneveld et al. Teaching and Teacher Education 118 (2022) 103826

Appendix B

Table: Gender Grading Bias Against Male students. Double Mediation of Implicit Gender-typed Associations and Gender-typed Expectations in the Relation Between
Teachers' Gender and Gender Grading Bias against Male Students

Dutch Language (N ¼ 131) Math (N ¼ 227)

(1) IA (2) IE (3) GGB (4) IA (5) IE (6) GGB

Constant 0.099 (0.066) 0.076 (0.089) 0.191 (0.045) 0.539 (0.041) 0.032 (0.064) 0.398 (0.062)
Gender Teacher 0.509** (0.075) 0.084 (0.118) 0.028 (0.059) 0.270** (0.055) 0.065 (0.067) 0.011 (0.066)
Implicit Associations 0.053 (0.119) 0.086 (0.060) 0.155* (0.077) 0.036 (0.076)
Expectations Talent 0.363** (0.044) 0.430** (0.065)
R2 0.262** 0.004 0.356** 0.097** 0.018 0.170**

Constant 0.099 (0.066) 0.210 (0.090) 0.167 (0.052) 0.539 (0.041) 0.194 (0.069) 0.370 (0.069)
Gender Teacher 0.509** (0.075) 0.192 (0.119) 0.012 (0.068) 0.270** (0.055) 0.048 (0.073) 0.020 (0.071)
Implicit Associations 0.104 (0.120) 0.080 (0.068) 0.163 (0.084) 0.114 (0.083)
Expectations Effort 0.243** (0.050) 0.072 (0.066)
R2 0.262** 0.020 0.171** 0.097** 0.017 0.012

Note: Standard Deviations in parentheses. IA ¼ Implicit Gender-typed Associations, IE ¼ Implicit Gender-typed Expectations, GGB ¼ Gender Grading Bias, ** p < 0.01, *
p < 0.05.

Table: Gender Grading Bias against Males Indirect Effects of Gender on Gender Grading Bias against Male Students

Dutch Language Math

Effect Size (SE) LLCI ULCI Effect Size (SE) LLCI ULCI

Total Mediated Effect 0.023 (0.050) 0.079 0.117 0.000 (0.033) 0.065 0.065
Mediation through IA 0.044 (0.031) 0.018 0.103 0.010 (0.019) 0.028 0.048
Mediation through IT 0.031 (0.041) 0.116 0.044 0.028 (0.031) 0.093 0.030
Double mediation 0.010 (0.020) 0.027 0.052 0.018 (0.011) 0.000 0.043

Total Mediated Effect 0.007 (0.045) 0.085 0.093 0.031 (0.025) 0.014 0.084
Mediation through IA 0.041 (0.035) 0.029 0.107 0.031 (0.024) 0.013 0.083
Mediation through IT 0.047 (0.029) 0.111 0.004 0.004 (0.008) 0.010 0.023
Double mediation 0.013 (0.017) 0.017 0.052 0.003 (0.004) 0.012 0.003

Note: Standard Deviations in parentheses. IA ¼ Implicit Gender-typed Associations, IE ¼ Implicit Gender-typed Expectations, GGB ¼ Gender Grading Bias, ** p < 0.01, *
p < 0.05.

Table: Gender Grading Bias Against Female students. Double Mediation of Implicit Gender-typed Associations and Gender-typed Expectations in the Relation Between
Teachers' Gender and Gender Grading Bias against Female Students

Dutch Language (N ¼ 131) Math (N ¼ 227)

(1) IA (2) IE (3) GGB (4) IA (5) IE (6) GGB

Constant 0.099 (0.066) 0.076 (0.089) 0.218 (0.044) 0.539 (0.041) 0.032 (0.064) 0.335 (0.059)
Gender Teacher 0.509** (0.075) 0.084 (0.118) 0.008 (0.058) 0.270** (0.055) 0.065 (0.067) 0.044 (0.062)
Implicit Associations 0.052 (0.119) 0.012 (0.058) 0.155* (0.077) 0.098 (0.072)
Expectations Talent 0.425** (0.043) 0.342** (0.062)
R2 0.262** 0.004 0.437** 0.097** 0.018 0.123**

Constant 0.099 (0.066) 0.210 (0.090) 0.226 (0.056) 0.539 (0.041) 0.194 (0.069) 0.370 (0.063)
Gender Teacher 0.509** (0.075) 0.192 (0.119) 0.007 (0.074) 0.270** (0.055) 0.048 (0.073) 0.015 (0.066)
Implicit Associations 0.104 (0.120) 0.014 (0.073) 0.163 (0.084) 0.025 (0.076)
Expectations Effort 0.193** (0.054) 0.122* (0.060)
R2 0.262** 0.020 0.094** 0.097** 0.017 0.020

Note: Standard Deviations in parentheses. IA ¼ Implicit Gender-typed Associations, IE ¼ Implicit Gender-typed Expectations, GGB ¼ Gender Grading Bias, ** p < 0.01, *
p < 0.05.

12
L. Doornkamp, L.D. Van der Pol, S. Groeneveld et al. Teaching and Teacher Education 118 (2022) 103826

Table: Gender Grading Bias against Females Indirect Effects of Gender on Gender Grading Bias against Female Students

Dutch Language Math

Effect Size (SE) LLCI ULCI Effect Size (SE) LLCI ULCI

Total Mediated Effect 0.018 (0.055) 0.124 0.090 0.035 (0.030) 0.092 0.022
Mediation through IA 0.006 (0.031) 0.055 0.069 0.027 (0.022) 0.072 0.016
Mediation through IT 0.036 (0.046) 0.128 0.055 0.022 (0.023) 0.066 0.026
Double mediation 0.011 (0.023) 0.033 0.059 0.014 (0.008) 0.000 0.032

Total Mediated Effect 0.020 (0.052) 0.126 0.082 0.006 (0.025) 0.057 0.042
Mediation through IA 0.007 (0.044) 0.080 0.093 0.007 (0.023) 0.054 0.038
Mediation through IT 0.037 (0.024) 0.088 0.005 0.006 (0.011) 0.012 0.031
Double mediation 0.010 (0.015) 0.016 0.046 0.005 (0.004) 0.016 0.000

Note: Standard Deviations in parentheses. IA ¼ Implicit Gender-typed Associations, IE ¼ Implicit Gender-typed Expectations, GGB ¼ Gender Grading Bias, ** p < 0.01, *
p < 0.05.

References Dasgupta, N. (2004). Implicit ingroup favoritism, outgroup favoritism, and their
behavioral manifestations. Social Justice Research, 17(2), 143e169. https://
doi.org/10.1023/B:SORE.0000027407.70241.15
Banaji, M. R., & Greenwald, A. G. (1995). Implicit Gender Stereotyping in judgments
De Boer, E., Janssen, F. J. J. M., Dam, M., & Van Driel, J. H. (2019). Ontwikkeling van
of fame. Journal of Personality and Social Psychology, 68(2), 181e198. https://
agency van docenten in opleiding: Een ecologisch perspectief. Pedagogische
doi.org/10.1037/0022-3514.68.2.181
Studies, 96(6), 354e377.
Bem, S. L. (1981). Gender schema theory: A cognitive account of sex typing. Psy-
Denessen, E., Hornstra, L., Van den Bergh, L., & Bijlstra, G. (2022). Implicit measures
chological Review, 88, 354e364.
of teachers' attitudes and stereotypes, and their effects on teacher practice and
Berg, P., Palmgren, O., & Tyrefors, B. (2020). Gender grading bias in junior high
student outcomes: A review. Learning and Instruction, 78, Article 101437.
school mathematics. Applied Economics Letters, 27(11), 915e919. https://doi.org/
https://doi.org/10.1016/j.learninstruc.2020.101437
10.1080/13504851.2019.1646862
Deutsch, F. M., & Saxon, S. E. (1998). The double standard of praise and criticism for
Biesta, G., Priestley, M., & Robinson, S. (2015). The role of beliefs in teacher agency.
mothers and fathers. Psychology of Women Quarterly, 22(4), 665e683. https://
Teachers and teaching, 21(6), 624e640. https://doi.org/10.1080/
doi.org/10.1111/j.1471-6402.1998.tb00184.x
13540602.2015.1044325
Devine, P. G., Forscher, P. S., Cox, W. T. L., Kaatz, A., Sheridan, J., & Carnes, M. (2017).
Blanton, H., Jaccard, J., Klick, J., Mellers, B., Mitchell, G., & Tetlock, P. E. (2009). Strong
A gender bias habit-breaking intervention led to increased hiring of female
claims and Weak Evidence: Reassessing the Predictive Validity of the IAT.
faculty in STEMM departments. Journal of Experimental Social Psychology, 73,
Journal of Applied Psychology, 94(3), 567e582. https://doi.org/10.1037/a0014665
211e215. https://doi.org/10.1016/j.jesp.2017.07.002
Breda, T., Jouini, E., Napp, C., & Thebault, G. (2020). Gender stereotypes can explain
DUO. (2020). 03 Lesgevend personeel voortgezet onderwijs (in personen). Con-
the gender-equality paradox. Proceedings of the National Academy of Sciences,
sulted from https://duo.nl/open_onderwijsdata/databestanden/vo/
117(49), 31063e31069. https://doi.org/10.1073/pnas.2008704117
onderwijspersoneel/vo-personeel3.jsp.
Breda, T., & Ly, S. T. (2015). Professors in core science fields are biased in favor of
Ellemers, N. (2018). Gender Stereotypes. Annual Review of Psychology, 69, 275e298.
women : evidence from France. American Economic Journal: Applied Economics,
https://doi.org/10.4324/9781003041870-5
7(4), 53e75.
Endendijk, J. J., Groeneveld, M. G., van Berkel, S. R., Hallers-Haalboom, E. T.,
Brown, C. S., & Stone, E. A. (2016). Gender Stereotypes and Discrimination. How
Mesman, J., & Bakermans-Kranenburg, M. J. (2013). Gender Stereotypes in the
Sexism impacts development. Advances in Child Development and Behavior, 50,
Family Context: Mothers, Fathers, and siblings. Sex Roles, 68(9e10), 577e590.
105e133. https://doi.org/10.1016/bs.acdb.2015.11.001
https://doi.org/10.1007/s11199-013-0265-4
Burgess, D., & Borgida, E. (1999). Who women are, who women should be:
Endendijk, J. J., Van Baar, A. L., & Dekovi c, M. (2020). He is a stud, she is a slut! A
Descriptive and Prescriptive Gender Stereotyping in Sex Discrimination. Psy-
meta-analysis on the continued existence of sexual double standards. Person-
chology, Public Policy, and Law, 5(3), 665e692. https://doi.org/10.1037/1076-
ality and Social Psychology Review, 24(2), 163e190. https://doi.org/10.1177/
8971.5.3.665
1088868319891310
Burgess, S., & Greaves, E. (2013). Test scores, subjective assessment, and stereo-
Espinoza, P., Are ^as da Luz Fontes, A. B., & Arms-Chavez, C. J. (2014). Attributional
typing of ethnic minorities. Journal of Labor Economics, 31(3), 535e576. https://
gender bias: Teachers' ability and effort explanations for students' math per-
doi.org/10.1086/669340
formance. Social Psychology of Education, 17(1), 105e126. https://doi.org/
Bussey, K., & Bandura, A. (1999). Social Cognitive Theory of Gender development
10.1007/s11218-013-9226-6
and Differentiation. Psychological Review, 106(4), 676e713.
Falch, T., & Naper, L. R. (2013). Educational evaluation schemes and gender gaps in
Caldwell, E., & Hartnett, R. (1967). Sex Bias in college Grading? Journal of Educational
student achievement. Economics of Education Review, 36, 12e25. https://doi.org/
Measurement, 4(3), 129e132.
pez, J. A., Nordfjell, O. B., Gaini, F., & Heikkinen, M. (2018). Promising 10.1016/j.econedurev.2013.05.002
Cardona Lo
Fazio, R. H., & Olson, M. A. (2003). Implicit Measures in Social Cognition Research:
Nordic practices in gender equality promotion: Developing teacher education
Their Meaning and use. Annual Review of Psychology, 54, 297e327. https://
dialogue, practice, and policy cycles on-line. Policy Futures in Education, 16(5),
doi.org/10.1146/annurev.psych.54.101601.145225
605e619. https://doi.org/10.1177/1478210317722286
Fiske, S. T. (2017). Prejudices in Cultural Contexts: Shared Stereotypes (Gender, Age)
Carnes, M., Devine, P. G., Baier Manwell, L., Byars-Winston, A., Fine, E., Ford, C. E.,
Versus Variable Stereotypes (race, ethnicity, religion). Perspectives on Psycho-
Forscher, P., Isaac, C., Kaatz, A., Magua, W., Palta, M., & Sheridan, J. (2015). The
logical Science, 12(5), 791e799. https://doi.org/10.1177/1745691617708204
effect of an intervention to break the gender bias habit for faculty at one
Forscher, P. S., Mitamura, C., Dix, E. L., Cox, W. T. L., & Devine, P. G. (2017). Breaking
institution: a cluster randomized, controlled trial. Academic Medicine Journal of
the prejudice habit: Mechanisms, timecourse, and longevity. Journal of Experi-
the Association of American Medical Colleges, 90(2), 221e230. https://doi.org/
mental Social Psychology, 72, 133e146. https://doi.org/10.1016/j.jesp.2017.04.009
10.1097/ACM.0000000000000552
Gansen, H. M. (2019). Push-Ups Versus clean-Up: Preschool Teachers' Gendered
CBS. (2022). Hoeveel mensen met een migratieachtergrond wonen in Nederland?
Beliefs, Expectations for Behavior, and disciplinary Practices. Sex Roles, 80,
Consulted from. https://www.cbs.nl/nl-nl/dossier/dossier-asiel-migratie-en-
393e408. https://doi.org/10.1007/s11199-018-0944-2
integratie/hoeveel-mensen-met-een-migratieachtergrond-wonen-in-
Gibbons, S., & Chevalier, A. (2008). Assessment and age 16þ education participation.
nederland-.
Research Papers in Education, 23(2), 113e123. https://doi.org/10.1080/
Charlesworth, T. E. S., & Banaji, M. R. (2021). Patterns of Implicit and Explicit Ste-
02671520802048638
reotypes III: Long-Term Change in Gender Stereotypes. Social Psychological and
Girod, S., Fassiotto, M., Grewal, D., Ku, M., Sriram, N., Nosek, B., & Valantine, H.
Personality Science, 13(1), 14e26. https://doi.org/10.1177/1948550620988425
(2016). Reducing Implicit Gender leadership Bias in Academic medicine With
Cole, M. S., Feild, H. S., & Giles, W. F. (2004). Interaction of recruiter and applicant
an Educational Intervention. Academic Medicine, 91(8), 1143e1150. https://
gender in resume evaluation: A field study. Sex Roles, 51(9e10), 597e608.
doi.org/10.1097/ACM.0000000000001099
https://doi.org/10.1007/s11199-004-5469-1
Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and Using the
Cornwell, C., Mustard, D. B., & Van Parys, J. (2013). Noncognitive skills and the
Implicit Association Test: I. An Improved Scoring Algorithm. Journal of Person-
gender disparities in test scores and teacher assessments: Evidence from pri-
ality and Social Psychology, 85(2), 197e216. https://doi.org/10.1037/0022-
mary school. Journal of Human Resources, 48(1), 236e264. https://doi.org/
3514.85.2.197
10.3368/jhr.48.1.236
Groeneveld, M. G., Van der Pol, L. D., De Vries, E. E., & Mesman, J. (2021). Triadic
Crisp, R. J., Bache, L. M., & Maitner, A. T. (2009). Dynamics of social comparison in
Family conversations About Gender: Children as driving forces and messiness in
counter-stereotypic domains: Stereotype boost, not stereotype threat, for
messages They Receive. Journal of Family Psychology. https://doi.org/10.1037/
women engineering majors. Social Influence, 4(3), 171e184. https://doi.org/
fam0000931
10.1080/15534510802607953
Haines, E. L., Deaux, K., & Lofaro, N. (2016). The Times They Are a-changing or Are

13
L. Doornkamp, L.D. Van der Pol, S. Groeneveld et al.
They Not? A Comparison of Gender Stereotypes, 1983e2014. Psychology of
Women Quarterly, 40(3), 353e363. https://doi.org/10.1177/0361684316634081
Hanna, R. N., & Linden, L. L. (2012). Discrimination in grading. American Economic
Journal: Economic Policy, 4(4), 146e168. https://doi.org/10.1257/pol.4.4.146
Hayes, A. F. (2018). Introduction to mediation, moderation, and conditional process
analysis: A regression-based approach. London: The Guilford Press.
Heller, K., Finsterwald, M., & Ziegler, A. (2001). Implicit theories of German Math-
ematics and physics Teachers on Gender Specific giftedness and Motivation.
Psychologische Beitra €ge, 43(1), 172.
Heyder, A., & Kessels, U. (2017). Boys don't Work? On the Psychological benefits of
Showing low Effort in High School. Sex Roles, 77, 72e85. https://doi.org/10.1007/
s11199-016-0683-1
Hinnerich, B. T., Ho €glin, E., & Johannesson, M. (2011). Are boys discriminated in
Swedish high schools? Economics of Education Review, 30(4), 682e690. https://
doi.org/10.1016/j.econedurev.2011.02.007
Hyde, J. S., Bigler, R. S., Joel, D., Tate, C. C., & Van Anders, S. M. (2019). The future of
sex and gender in psychology: Five challenges to the gender binary. American
Psychologist, 74(2), 171e193. https://doi.org/10.1037/amp0000307
Jones, S., & Myhill, D. (2004). Troublesome boys” and “compliant girls”: Gender
identity and perceptions of achievement and underachievement. British Journal
of Sociology of Education, 25(5). https://doi.org/10.1080/0142569042000252044
Keiser, L. R., Wilkins, V. M., Meier, K. J., & Holland, C. A. (2002). Lipstick and loga-
rithms: Gender, Institutional Context, and representative bureaucracy. American
Political Science Review, 96(3), 553e564. https://doi.org/10.1017/
S0003055402000321
Kinias, Z, & Kim, H. S. (2011). Culture and gender inequality: psychological conse-
quences of perceiving gender inequality. Group Process. Intergr. Relat., 15(1),
89e103. https://doi.org/10.1177/1368430211408940. In this issue.
Lavy, V. (2008). Do gender stereotypes reduce girls' or boys' human capital out-
comes? Evidence from a natural experiment. Journal of Public Economics,
92(10e11), 2083e2105. https://doi.org/10.1016/j.jpubeco.2008.02.009
Lavy, V., & Sand, E. (2018). On the origins of gender gaps in human capital: Short-
and long-term consequences of teachers' biases. Journal of Public Economics, 167,
263e279. https://doi.org/10.1016/j.jpubeco.2018.09.007
Lenton, A. P., Bruder, M., & Sedikides, C. (2009). A meta-analysis on the malleability
of automatic gender stereotypes. Psychology of Women Quarterly, 33(2),
183e196. https://doi.org/10.1111/j.1471-6402.2009.01488.x
Liben, L. S., & Bigler, R. S. (2002). The developmental course of gender differentia-
tion: Conceptualizing, measuring, and evaluating constructs and pathways.
Monographs of the Society for Research in Child Development, 67(2), 1e183.
Lindahl, E. (2016). Are teacher assessments biased? e evidence from Sweden. Ed-
ucation Economics, 24(2), 224e238. https://doi.org/10.1080/
09645292.2015.1014882
Lyness, K. S., & Heilman, M. E. (2006). When fit is fundamental: Performance
evaluations and promotions of upper-level female and male managers. Journal
of Applied Psychology, 91(4), 777e785. https://doi.org/10.1037/0021-
9010.91.4.777
Martin, C. L., & Dinella, L. M. (2012). Congruence between gender stereotypes and
activity preference in self-identified tomboys and non-Tomboys. Archives of
Sexual Behavior, 41(3), 599e610. https://doi.org/10.1007/s10508-011-9786-5
Matej
u, P., & Smith, M. L. (2015). Are boys that bad? Gender gaps in measured skills,
grades and aspirations in Czech elementary schools. British Journal of Sociology
of Education, 36(6), 871e895. https://doi.org/10.1080/01425692.2013.874278
Muntoni, F., & Retelsdorf, J. (2018). Gender-specific teacher expectations in read-
ingdThe role of teachers' gender stereotypes. Contemporary Educational Psy-
chology, 54, 212e220. https://doi.org/10.1016/j.cedpsych.2018.06.012
Nosek, B. A., Banaji, M. R., & Greenwald, A. G. (2002a). Harvesting implicit group
attitudes and beliefs from a demonstration web site. Group Dynamics, 6(1),
101e115. https://doi.org/10.1037/1089-2699.6.1.101
Nosek, B. A., Banaji, M. R., & Greenwald, A. G. (2002b). Math ¼ male, me ¼ female,
therefore math s me. Journal of Personality and Social Psychology, 83(1), 44e59.
https://doi.org/10.1037/0022-3514.83.1.44
14
Teaching and Teacher Education 118 (2022) 103826
Nosek, B. A., Greenwald, A. G., & Banaji, M. R. (2005). Understanding and Using the
Implicit Association Test: II. Method Variables and construct Validity. Person-
ality and Social Psychology Bulletin, 31(2), 166e180. https://doi.org/10.1177/
0146167204271418
Nosek, B. A., Smyth, F. L., Sriram, N., Lindner, N. M., Devos, T., Ayala, A., et al. (2009).
National Differences in Gender-Science Stereotypes predict National Sex Dif-
ferences in Science and Math Achievement. Proceedings of the National Academy
of Sciences, 106(26), 10593e10597. https://doi.org/10.1073/pnas.0809921106
Nürnberger, M., Nerb, J., Schmitz, F., Keller, J., & Sütterlin, S. (2016). Implicit Gender
Stereotypes and essentialist Beliefs predict preservice Teachers tracking rec-
ommendations. Journal of Experimental Education, 84(1), 152e174. https://
doi.org/10.1080/00220973.2015.1027807
Obioma, I. F., Hentschel, T., & Hernandez Bark, A. S. (2021). Gender stereotypes and
self-characterizations in Germany and Nigeria: A cross-cultural comparison.
Journal of Applied Social Psychology, 1e17. https://doi.org/10.1111/jasp.12801, 00.
Powell, G. N. (1986). Applicants' qualifications on recruiters' evaluations. Psycho-
logical Reports, 58, 1003e1010. https://doi.org/10.2466/pr0.1986.58.3.1003
Priestley, M., Biesta, G. J. J., & Robinson, S. (2015). Teacher agency: what is it and why
does it matter? In R. Kneyber, & J. Evers (Eds.), Flip the system: Changing edu-
cation from the bottom up. London: Routledge.
Protivínský, T., & Münich, D. (2018). Gender Bias in teachers' grading: What is in the
grade. Studies in Educational Evaluation, 59, 141e149. https://doi.org/10.1016/
j.stueduc.2018.07.006
Riegle-Crumb, C., & Humphries, M. (2012). Exploring Bias in Math Teachers' Per-
ceptions of Students' Ability by Gender and race/ethnicity. Gender and Society,
26(2), 290e322. https://doi.org/10.1177/0891243211434614
Rudman, L. A., & Kilianski, S. E. (2000). Implicit and explicit attitudes toward female
authority. Personality and Social Psychology Bulletin, 26(11), 1315e1328. https://
doi.org/10.1177/0146167200263001
Sagebin Bordini, G., & Sperb, T. M. (2013). Sexual Double Standard: A Review of the
Literature Between 2001 and 2010. Sexuality and Culture, 17(4), 686e704.
https://doi.org/10.1007/s12119-012-9163-0
Schmenk, B. (2004). Language Learning : A Feminine Domain ? The Role of Ster-
eotyping in Constructing Gendered Learner identities. TESOL Quarterly, 38(3),
514e524. https://doi.org/10.2307/3588352
Schussler, D. L., Stooksberry, L. M., & Bercaw, L. A. (2010). Understanding Teacher
candidate dispositions: Reflecting to build Self-Awareness. Journal of Teacher
Education, 61(4), 350e363. https://doi.org/10.1177/0022487110371377
Siegle, D., & Reis, S. M. (1998). Gender Differences in Teacher and Student Percep-
tions of Gifted Students' Ability and Effort. Gifted Child Quarterly, 42(1), 39e47.
https://doi.org/10.1177/001698629804200105
Smeding, A. (2012). Women in Science, technology, Engineering, and Mathematics
(STEM): An Investigation of Their Implicit Gender Stereotypes and Stereotypes'
connectedness to Math Performance. Sex Roles, 67(11e12), 617e629. https://
doi.org/10.1007/s11199-012-0209-4
Stobart, G., Elwood, J., & Quinlan, M. (1992). Gender Bias in Examinations : How
Equal Are the opportunities. British Educational Research, 18(3), 261e276.
https://www.jstor.org/stable/1500831.
Tabachnick, B. G., & Fidell, L. S. (2012). Using multivariate statistics (6th ed.). Harper
Collins.
Van den Bergh, L., Denessen, E., Hornstra, L., Voeten, M., & Holland, R. W. (2010). The
implicit prejudiced attitudes of teachers: Relations to teacher expectations and
the ethnic achievement gap. American Educational Research Journal, 47(2),
497e527. https://doi.org/10.3102/0002831209353594
Wulff, J. N., & Villadsen, A. R. (2019). Are survey Experiments as valid as Field Ex-
periments in Management Research? An Empirical Comparison Using the Case
of Ethnic employment Discrimination. European Management Review, 17(1),
347e356. https://doi.org/10.1111/emre.12342
Xu, X., & Meier, K. J. (2021). Separating symbolic and active representation: a mixed
methods study of gender and education in China. Public Management Review,
1e23. https://doi.org/10.1080/14719037.2021.1900352, 00(00.